C[omp]ute

From: andrew cooke <andrew@...>

Date: Thu, 30 Jun 2011 19:38:39 -0400

This is work-related; I'm posting it to my personal blog just to make it
available to others.  Credit to Joel Edwards at USGS Albuquerque (who
explained most of this to me) and ISTI (my employers).


The Q330 has 8 output (control) lines that can be connected to local hardware
(nominally 4 lines per sensor).  These use 0/5V levels and are controlled
directly by the "Sensor Control Bitmap".

An invocation of the C1_QCAL command can set these lines "however it pleases".
For example, if the "Sensor Control Bitmap" is set to 0x1 then the first line
is set to 5V and all others are set to 0V.


So far, so simple.

However, the Q330 documentation and firmware also support a set of conventions
that allow the meaning of these lines to be stored and read, and so allow the
control software to make "intelligent" decisions about how to set the lines
for particular commands.

These conventions are as follows:

  - A set of names for the 8 lines.  These are displayed in the Willard
    software, named as:

      Sensor A:
	GENEN 1-A, GENEN 2-A, GENEN 3-A, CALEN-A

      Sensor B:
	GENEN 1-B, GENEN 2-B, GENEN 3-B, CALEN-B

  - A set of "labels" that identify uses for the lines:

      Not doing calibration or re-centring 
      Sensor A calibration 
      Sensor A centring 
      Sensor A capacitive coupling 
      Sensor B calibration 
      Sensor B centring 
      Sensor B capacitive coupling 
      Sensor A Lock 
      Sensor A Unlock 
      Sensor A Aux1 
      Sensor A Aux2 
      Sensor B Lock 
      Sensor B Unlock 
      Sensor B Aux1 
      Sensor B Aux2 

The firmware allows each name to be associated with (1) a label and (2) a flag
for active high.  This is written to the Q330 via the C1_SSC command and read
via C1_SC.

So, for example, the user might decide that Lines GENEN 1-A and CALEN-A are
both labelled "Sensor A calibration", and that the first of these is "active
high" while the second is "active low".


Given the above we can now describe how control software uses the conventions
to make "intelligent decisions".  For example, when running calibration the
software "knows" that it must trigger calibration for sensor A.  Consulting
the table of information described above it can infer that it must set GENEN
1-A (bit 0) to 1 (active high) and CALEN-A (bit 3) to 0 (active low).  The
other bits would need to be set to their "inactive" states (assuming they were
set to labels that had nothing to do with calibration).


In summary:

 - The Q330 itself does not apply any logic itself.  It simply sets the line
   voltages according to the bitmap supplied.
 - A set of conventions allow the physical configuration (wiring) to be
   described and saved on the Q330.
 - Control software can read this information and, following the conventions,
   control the output lines correctly.


Finally, the "error" in qscal comes from no line being labelled with
"capacitive coupling".  The software checks for this when a capacitively
coupled calibration is made.  This is a "logical" check (if the operator has
not recorded any wire to be used to enable capacitive coupling then it probably
does not make sense to ask for such a thing) inferred from the conventions
above.  It does not reflect any actual error in the low level behaviour of the
system.